OpenAI claims $1 million prize for solving Navier-Stokes problem, faces scrutiny

OpenAI’s language model has produced a proof that it claims settles the Navier‑Stokes existence and smoothness problem, and the claim could make the company eligible for the Clay Mathematics Institute’s $1 million Millennium Prize – but only by invoking a controversial loophole in the problem’s wording.

The Navier‑Stokes equations describe fluid flow and are central to physics, engineering and climate science. In 2000 the Clay Institute listed the existence and smoothness of solutions in three‑dimensional space as one of seven “Millennium Prize Problems,” offering a cash reward to anyone who could prove that solutions always exist and remain smooth, or provide a counterexample. No mathematician has yet succeeded, and the problem has become a benchmark for deep analytical skill.

OpenAI’s proof, generated by a GPT‑4‑based system, sidesteps the traditional formulation by restricting the equations to a periodic domain on a torus and assuming a specific type of weak solution that avoids the singularities the original problem seeks to exclude. The approach satisfies the literal text of the prize’s criteria, which refer to “the Navier‑Stokes equations on ℝ³” but do not explicitly forbid such domain reductions. Several mathematicians have warned that the loophole undermines the spirit of the challenge. Professor Terence Tao called the result “interesting as a demonstration of AI‑assisted reasoning, but not a solution to the problem as originally posed.” Others note that the proof still requires rigorous peer review and verification of the AI’s logical steps, a process that could take months.

The next step will be a formal submission to the Clay Institute, where a panel of experts will assess whether the proof meets the intended standards. If the panel accepts the loophole, OpenAI could claim the prize; if not, the work may still influence how AI tools are used in mathematical research. Regardless of the outcome, the episode highlights the growing role of artificial intelligence in tackling problems once thought to be the exclusive domain of human insight.